AI in the Staffroom: Getting Your Department Onboard with AI Planning

Every department has one teacher who's brilliant with AI and three who've decided it's a fad, one who'd love training and two who fear it means their subject expertise is being replaced. Bringing a whole department onto AI planning isn't a software rollout; it's a change-management exercise with humans who have excellent reasons for their scepticism. Done badly, you get a policy nobody follows and a quiet spreadsheet of AI-generated notes nobody checks. Done well, you get the biggest collective upgrade to planning quality in a generation. Here's the staff-room playbook.

Start with the sceptics' best argument

The fastest way to lose a staffroom is to open with enthusiasm. Open instead with the strongest sceptical point, said honestly: most AI-generated text is generic, alignment can't be trusted blindly, and there are real risks in outsourcing professional judgement. Then add the part the sceptics don't usually know: the tools are designed so none of that happens when they're used professionally. The planner plans; the teacher verifies. A department that hears its own doubts spoken first will listen to what follows. A department that hears enthusiasm will stop listening.

The demonstration that actually converts

Forget the feature tour. The converting demonstration is a head-to-head with the department's real work: take next week's actual topics, generate notes live, then put the AI drafts beside the department's existing notes and critique them together. The result surprises people in both directions: the drafts are more structured than expected, and their errors (a wrong example, a misaligned objective, a timing that doesn't sum) are exactly the things a professional catches in minutes. The teachers who fear replacement see that the errors are professional-level, meaning the value stays professional. The enthusiasts see the limits, which makes them safer users. Everyone leaves with the same model: draft, verify, own.

Agree the professional standards first, the tool second

Before adoption spreads, the department needs its own lines, written by the teachers themselves. Ours read like this in most departments that do this well: the note that reaches a classroom has been read and edited by a human teacher; objectives must trace to the actual indicator, verified by the teacher; no student data goes into any tool that doesn't protect it; and the reflection is always written after teaching, never generated before it. A page of department-owned standards beats a circular from above, because the people enforcing the standards are the people who wrote them.

Training that respects expertise

The training session that works is ninety minutes, hands-on, and built around the work teachers were going to do anyway: bring next week's scheme rows, plan the actual lessons, in the room, with help. Two rules make it land. First, nobody generates anything without writing a brief first (who's in the room, what exactly is being taught, what the lesson must survive); the brief is the professional skill, and teaching it IS the training. Second, every teacher leaves having produced at least one note they'd actually teach, because the first personal success does more for adoption than any amount of policy. For the prompt craft itself, see prompting your AI lesson planner.

The co-planning dividend

Here's where AI planning becomes a departmental superpower rather than a personal one. When everyone plans from the same aligned drafts, co-planning changes shape: instead of two teachers each writing from scratch, one reviews the other's generated-and-edited notes against the department's standards, in half the time, at higher quality. Subject banks build themselves as a by-product: every verified note is filed by indicator, and the bank compounds. Within a year, a department that adopts this has a scheme's worth of professional notes that new staff inherit, exam classes draw from, and everyone's evenings thank them for.

Handle the workload question honestly

The sceptic's deepest question is "who benefits?" If AI planning means everyone plans more, faster, for the same total free time, it's a treadmill. Say plainly what the time is for: reclaimed evenings, more feedback, less Sunday. Heads who adopt AI planning and simultaneously raise output expectations kill the goodwill in one term. The departments that thrive write the reinvestment down: the hours saved go to feedback and rest, visibly, as policy. That's also the answer to the teacher who asks whether this is just a way to increase class sizes.

The first term, step by step

Week one: the head-to-head demonstration and the standards discussion. Week two: the hands-on session, real lessons planned. Weeks three to six: each teacher uses the planner for their own notes, with a fortnightly twenty-minute share where two volunteers show a draft and how they edited it. Week eight: the department reviews what changed (time spent, note quality, student-facing results) and decides the next step together. The volunteering rate at week nine tells you everything; departments that run this sequence typically arrive at near-full adoption within a term, not because of policy, but because the early adopters' evenings got measurably better and the sceptics noticed.

The department that does this well hasn't adopted software. It has adopted a workflow where the machine does the assembly, the professionals do the judgement, and the standard of every note in the department rises together. That's the actual prize, and it's worth one well-run staff meeting to start. For the workload system around it, read cutting teacher workload without cutting corners, and for the alignment engine underneath, see the Curriculum Engine.

What to do when someone opts out entirely

There will be one. A capable teacher who wants nothing to do with any of it. The mistake is forcing compliance (you get resentful, unverified generation, the worst of both worlds) and the other mistake is pretending the question isn't being asked by three silent colleagues. The working middle: hold them to the department's output standards, never to the tool. If their notes are aligned, verified and on time, produced however they like, the department has no case. Make that explicit at the start, and watch what happens: within a term, most opt-outs quietly adopt the workflow for the most human of reasons. They notice their planned-free colleagues going home on time.